Editorial correction — September 20, 2026. The earlier article treated different survey measures and periods as interchangeable. Gallup also found rising attendance. PRRI measured weekly attendance over a different interval; neither result should erase the other.
Two surveys, two definitions
Gallup’s April 16, 2026 analysis compared pooled 2022–2023 interviews with 2024–2025. Among men ages 18–29, the share describing religion as very important rose from 28% to 42%. Attendance at least monthly also increased, from about 33% to 40%. It is incorrect to describe Gallup as finding an attitude change with no attendance change.
PRRI’s 2025 Census of American Religion compared 2024 with 2025 and found weekly attendance among young men unchanged at 20%. Weekly attendance is a narrower category than monthly-or-more attendance. Annual comparisons also differ from comparisons between pooled two-year periods.
How apparently conflicting findings can coexist
Imagine someone who previously attended once a year and begins attending once a month. They would move into an at-least-monthly category without moving into a weekly category. This illustration explains why the measures are not interchangeable; it does not establish what any particular survey respondent did.
Likewise, a rise that occurred before the final year of a two-year interval could appear in a pooled comparison while a later one-year comparison remains flat. Different samples and survey methods add another reason to compare the questions before declaring one survey wrong.
Belief, membership and conduct
My interest in religious institutions is not satisfied by a winning headline about revival or decline. Stated importance, affiliation, attendance and service to neighbors are distinct questions. A congregation can count attendance accurately while still needing to examine whether its members care for people outside the room.
The ethical test I would put to a faith community is concrete: does renewed involvement produce sustained responsibility, honest leadership and practical care? Neither survey answers that question. It belongs to the people who participate and the neighbors affected by their conduct.
When discussing a religious trend, write down the population, dates, exact attendance threshold and source before drawing a conclusion. These results deserve attention on their own terms. They do not justify calling young men hypocritical, declaring young women morally deficient, or presenting a national spiritual verdict as a statistical finding.